ICYMI Video: Blumenthal Highlights Dangers of Artificial Intelligence in Spreading Disinformation with Election Deepfakes

[WASHINGTON, D.C.] – In case you missed it, yesterday, U.S. Senator Richard Blumenthal (D-CT), Chair of the Senate Judiciary Subcommittee on Privacy, Technology, and the Law, raised significant concerns regarding the threat AI poses to elections through the use of deepfakes during a hearing titled “Oversight of AI: Election Deepfakes.”

“We are here today at this Subcommittee meeting because a deluge of deception, disinformation, and deepfakes are about to descend on the American public. The form of their arrival will be political ads, and other forms of disinformation that are made possible by artificial intelligence,” said Blumenthal. “There is a clear and present danger to our democracy. This world of disinformation doesn’t have to be our future. We have agency, we can take action. And we are here today not only to hear about the dangers, but also to look forward to action that we can take in the United States Congress.”

This hearing comes after deepfakes were used to impersonate President Biden to discourage voters from participating in the New Hampshire presidential primary in January.

“Artificial intelligence is already being used to interfere with our elections, sowing lies about candidates and suppressing the vote. We already have a chilling example. This January, thousands of New Hampshire residents received a call impersonating President Biden, telling them not to vote—not to vote—in the state’s primary,” said Blumenthal. “Between the ease of use, and the increasing interest from foreign adversaries and domestic political interests, our democracy is facing a perfect storm.”

To show how realistic deepfakes are, Blumenthal ran an experiment in which he and his colleagues were asked to determine which recording of his voice was AI-generated. Ben Colman, CEO and Co-Founder of Reality Defender and a witness at the hearing, helped organize the experiment and revealed that although Blumenthal had earlier recorded the same lines that were read during the hearing, the audio that they played were all deepfakes.

“And as you guys think about which ones are real and fake, we’re going to share with you the surprise that they’re actually all fake. The challenge and the opportunity is that anybody with a Google search and internet connection can make something as entertaining or as dangerous as they can imagine,” said Colman.

Video of Blumenthal’s opening remarks can be found here, and video of the deepfake experiment can be found here. Transcript of Blumenthal’s opening remarks can be found below.

We are here today at this Subcommittee meeting because a deluge of deception, disinformation, and deepfakes are about to descend on the American public. The form of their arrival will be political ads, and other forms of disinformation that are made possible by artificial intelligence. There is a clear and present danger to our democracy. This world of disinformation doesn’t have to be our future. We have agency, we can take action. And we are here today not only to hear about the dangers, but also to look forward to action that we can take in the United States Congress.

But we should make no mistake—the threat of political deepfakes is real, it’s happening now, it’s not science fiction coming at some point in the future, possibly or hypothetically. Artificial intelligence is already being used to interfere with our elections, sowing lies about candidates and suppressing the vote.

We already have a chilling example. This January, thousands of New Hampshire residents received a call impersonating President Biden, telling them not to vote—not to vote—in the state’s primary. It is important for the American people to hear exactly what was said.

[New Hampshire Biden deepfake call] What a bunch of malarkey. You know the value of voting democratic when our votes count. It’s important that you save your vote for the November election. We’ll need your help in electing Democrats up and down the picket. Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again. Your vote makes a difference in November, not this Tuesday. If you would like to be removed from future calls, please press 2 now.

It's important for the American people to hear what impersonation and deepfakes look like. It’s also important to know that that’s what suppression of voter turnout looks like.

The deepfake of President Biden wasn’t made by a computer whizz, or some computer science graduate student, or anybody with any particular skill. It was made by a street magician, whose previous claim to fame was that he has world records in spoon bending and escaping straightjackets.

The voice cloning technologies used in that call were inconceivable just a few years ago. Now, they are free, online. Available to everyone.

And it’s not just voice cloning. Deepfake images and videos are disturbingly easy for anyone to create. 

Protecting our elections isn’t about Democrats versus Republican. Already, deepfakes have targeted candidates from across the political spectrum, and no one, literally no candidate, no voter, no one, is safe from them.

And if a street magician can cause this much trouble, imagine what Vladimir Putin or China can do. In fact, they’re doing it. National security officials and law enforcement have been shouting from the rooftops, as well as in our classified briefings, their fears about AI and foreign disinformation.

It’s happening. It’s here. Earlier this month, Microsoft revealed that social media accounts linked to the Chinese Communist Party were using AI to meddle in American politics. China has been caught using deepfakes to impersonate Americans to sow division and conspiracies theories—such as deepfake images to push the lie that the United States military caused wildfires in Hawaii.

Between the ease of use, and the increasing interest from foreign adversaries and domestic political interests, our democracy is facing a perfect storm.

When the American people can no longer recognize fact from fiction, quite literally, it will be impossible to have a democracy.

As we discussed in our last hearing, these deepfakes and rampant disinformation are also happening at time when local journalism is hanging by a thread.

Deepfakes have targeted not only Presidential candidates, but also Senate campaigns, and local elections, like the recent Chicago mayoral election.

Anyone can do it—even in the tiniest race. In some ways, local elections present an even bigger risk.

A deepfake of President Biden will attract national attention. It will be publicized as disinformation and deception. But deepfakes on a local election, state legislative contest, or city council—probably not. And, when a local newspaper is closed or understaffed, there may be no one doing fact-checking. No one to issue those Pinocchio images. And no one to correct the record. That’s a recipe for toxic and destructive politics.

Congress has the power—indeed the obligation—to stop this AI nightmare. There are common sense, bipartisan bills ready-to-go right now. I’m supporting them, a number of my colleagues have offered and supported them as well. Senators Klobuchar and Hawley’s legislation to prohibit deceptive political deepfakes, the Protect Elections from Deceptive AI Act. Requiring consent and watermarks for deepfakes, such as Senator Coons and Senator Blackburn’s, called the NO FAKES Act. On Section 230, Senator Hawley and I have a bill to ensure that there’s no question that Section 230 does not apply to AI, we can hold social media companies and Big Tech accountable for election deepfakes and other AI-driven harms.  

If we leave them unchecked, deepfakes and political deceptions will sow the seeds of our destruction as a democracy. It may sound like an exaggeration, but it is dangerously true.

And so, this world of disinformation and poisonous lies doesn’t have to be our future. Today, we need to begin to continue the process of making sure it isn’t our future. And with that, I’ll turn to the Ranking Member.

-30-